on 2/19/2008 2:09 PM Hugh E Cruickshank spake the following:
Hi All:
Over the weekend I install all the outstanding updates for our CentOS 4 based server. Since I had been holding off on these until I had addressed some disk space issues there were a large number (300+). I know my bad! After installing the updates I rebooted the system and it took forever to boot and once up there were problems connecting to some of our SAMBA shares. I checked the messages log file and found a multitude of entries similar to:
Feb 17 19:46:18 fisds0 named[23187]: nss_ldap: failed to bind to LDAP server 127.0.0.1: Can't contact LDAP server Feb 17 19:46:18 fisds0 named[23187]: nss_ldap: reconnecting to LDAP server...
These were being reported for named, nscd, smbd, statd, rquotad, etc.
I did some google searching and found some references to the change of the default value for the "bind_policy" parameter in the "ldap.conf" file from "soft" to "hard". I added and explicit "bind_policy soft" to the "/etc/ldap.conf" file and that has improved things dramatically.
However the errors are still being reported in the messages log file. The errors are valid as we do not have and LDAP server (on my list for a future project). What I am trying to figure out is why it is looking for one. I have done some additional google searching but I have not found any definitive answers. From what I have seen I suspect that the problem lays with our /etc/nsswitch.conf file and that I need to change references to "file ldap" to just "files".
I am loath to make such changes without something more definite then my personal suspicion. Can someone confirm that I am on the right track or, if not, point me in the right direction?
TIA
Regards, Hugh
As long as you have ldap entries in nsswitch.conf those services will attempt to look for ldap.